systemd: update to version 258

[virtual-designer]

Hi there,

This PR updates systemd to version 258, which is the latest as of now.

[fishilico]

Hello,

Thanks for your Pull Request.

As you see in the CI output, this update caused a regression in dbus-selinux:

2025-09-17T17:38:01.4816994Z ../dbus/dbus/dbus-sysdeps-unix.c:98:10: fatal error: systemd/sd-daemon.h: No such file or directory
2025-09-17T17:38:01.4817771Z    98 | #include <systemd/sd-daemon.h>
2025-09-17T17:38:01.4818174Z       |          ^~~~~~~~~~~~~~~~~~~~~
2025-09-17T17:38:01.4818554Z compilation terminated.

Did you manage to fix it?

[virtual-designer]

Hello,

Thanks for your Pull Request.

As you see in the CI output, this update caused a regression in dbus-selinux:

2025-09-17T17:38:01.4816994Z ../dbus/dbus/dbus-sysdeps-unix.c:98:10: fatal error: systemd/sd-daemon.h: No such file or directory
2025-09-17T17:38:01.4817771Z    98 | #include <systemd/sd-daemon.h>
2025-09-17T17:38:01.4818174Z       |          ^~~~~~~~~~~~~~~~~~~~~
2025-09-17T17:38:01.4818554Z compilation terminated.

Did you manage to fix it?

Hi,

Sorry for the stupid mistake. It was because of the way I wrote the copy command for header files, during build which would copy them into systemd-libs/usr/include/include. I wasn't careful enough.

I checked the CI logs again, turns out we have another error this time:

Run sudo /tmp/boots/usr/bin/arch-chroot /tmp/arch /bin/bash -ex -c \
+ ln -sfv /usr/share/zoneinfo/UTC /etc/localtime
'/etc/localtime' -> '/usr/share/zoneinfo/UTC'
+ hwclock --systohc
+ sed -i s/#en_US.UTF-8/en_US.UTF-8/ /etc/locale.gen
+ locale-gen
Generating locales...
  en_US.UTF-8... done
Generation complete.
+ echo LANG=en_US.UTF-8
+ echo qemu-arch-selinux
+ echo -e '127.0.0.1 localhost\n::1 localhost'
+ echo -e '[Match]\nName=en*\n[Network]\nDHCP=ipv4'
+ systemctl enable systemd-networkd.service
Failed to connect to system scope bus via local transport: No data available
Error: Process completed with exit code 1.

This is strange because I am using the package right now on machine. It's working. (I use CachyOS)

[fishilico]

Sorry for the stupid mistake. It was because of the way I wrote the copy command for header files, during build which would copy them into systemd-libs/usr/include/include. I wasn't careful enough.

I have not understood the issue, and I see systemd's PKGBUILD is still using mv commands: https://gitlab.archlinux.org/archlinux/packaging/packages/systemd/-/blob/main/PKGBUILD#L266-267

  mv "$pkgdir"/usr/include systemd-libs/include
  mv "$pkgdir"/usr/share/man/man3 systemd-libs/man3

Also, I see there are some very recent bugfixes to package systemd in https://gitlab.archlinux.org/archlinux/packaging/packages/systemd/-/commits/main :

Usually, when I see such activity, I prefer to delay a bit the update, to wait for a more stable version.

Regarding the CI issue with systemd-networkd.service, I will try running the CI again to see whether it was a transient issue (in GitHub Actions infrastructure) or something which really broke with systemd-selinux update.

[virtual-designer]

Sorry for the stupid mistake. It was because of the way I wrote the copy command for header files, during build which would copy them into systemd-libs/usr/include/include. I wasn't careful enough.

I have not understood the issue, and I see systemd's PKGBUILD is still using mv commands: https://gitlab.archlinux.org/archlinux/packaging/packages/systemd/-/blob/main/PKGBUILD#L266-267

  mv "$pkgdir"/usr/include systemd-libs/include
  mv "$pkgdir"/usr/share/man/man3 systemd-libs/man3

When I was building systemd, multiple times, especially due to a few build failures, what happened is that systemd-libs/include was already created, and then the mv command would copy the include directory inside systemd-libs/include, which is the same issue that we just dealt with. That's why I changed those commands to make sure even if those directories exist, it copies the files correctly. For CI/clean builds, this probably doesn't change anything. I can always change it back if you think that's what should be done.

Edit: You know what, just leave it. I have reverted that change.

[virtual-designer]

Additionally, the recent CI logs indicate there's a problem with the GNULib repository. I couldn't clone it locally either, seems like that server went out of memory due to some sort of request flooding or DDoS.

  -> Cloning findutils git repo...
Cloning into bare repository '/startdir/findutils-selinux/findutils'...
  -> Cloning gnulib git repo...
Cloning into bare repository '/startdir/findutils-selinux/gnulib'...
remote: fatal: Out of memory, malloc failed (tried to allocate 99729 bytes)        
remote: aborting due to possible repository corruption on the remote side.
fatal: protocol error: bad pack header
==> ERROR: Failure while downloading gnulib git repo
    Aborting...
Error: Process completed with exit code 1.